Chapter 2, Section 3
Part of AM Radio Facts and Trivia: The AM Broadcast Band in the USA. Updated November 15, 2026.
How AM broadcasting went from wires on rooftops to the complex phased arrays that shape every signal on the dial today.
When AM broadcasting began around 1920, stations borrowed antenna technology from maritime radiotelegraphy: horizontal long-wire antennas. These took two forms: the "flattop" (parallel wires on cross-arms) and the "cage" (wires in a cylindrical arrangement). Both were typically suspended between two masts or between a mast and a tall building, fed at the center (T-antenna) or at one end (L-antenna). Most early stations operated from rooftops in downtown buildings, with their studios, transmitters, and antennas all in the same location. The FCC logo still incorporates the old flattop antenna and its feeder system. What engineers didn't realize at first was that in most of these installations, the vertical feed wire did most of the actual radiating: the horizontal wires on top were essentially just capacitive loading.
In 1924, Stuart Ballantine published two landmark papers that changed everything. He showed mathematically that a vertical monopole antenna over a ground plane had a radiation resistance that increased with height up to a half wavelength, and that maximum horizontal (groundwave) radiation occurred at about 0.625 wavelength (225 electrical degrees). This meant a single vertical tower (properly insulated at its base and fed against a ground system) could outperform a two-mast flattop while being cheaper to build and requiring far less land. In 1931, New York's WABC (now WHSQ/WCBS's predecessor) erected what is believed to be the world's first medium-wave insulated-base vertical radiator in Paterson, New Jersey, and it made the cover of Radio-Craft magazine. By 1930-31, the Blaw-Knox company had patented its distinctive diamond cantilever tower design, one of the first purpose-built mast radiators. A 1942 Blaw-Knox ad claimed 70% of all US radio towers were their product. The era of the wire antenna was over.
As the number of AM stations exploded through the 1930s (from about 500 to over 2,000 by 1950), co-channel interference became severe, especially at night when skywave propagation could carry signals thousands of miles. The solution was the directional antenna: multiple towers fed with carefully controlled current amplitudes and phases to create radiation patterns that could be shaped to protect other stations. The first known directional antenna was built in 1927 for WFLA/WSUN in Clearwater, Florida, designed by Raymond Wilmotte using lumped-constant phasing networks. By 1937, 39 of the 646 AM stations on the air were using directional antennas. In 1935, WOR in New York built a landmark array designed by Bell Labs: two self-supporting towers with a taut cable and drop-wire as a third element. The 1941 North American Regional Broadcasting Agreement (NARBA) reshuffled the entire AM band, moved hundreds of stations to new frequencies, and established the Class system (I through IV) that required many stations to directionalize to protect clear-channel stations. This drove the massive adoption of directional arrays, and the tower farms we see today.
After NARBA and through the postwar boom, the number of AM stations tripled from about 2,000 to 4,000 by 1970, virtually all made possible by directional antenna technology. Today, the United States has more directional AM antenna systems than all other countries combined. Arrays range from simple two-tower patterns to the 12-tower monster at KFXR in Dallas. Tower heights are measured in electrical degrees: 90° is a quarter wavelength (the most common), while some stations use towers over 200° for maximum groundwave performance. Sectionalized towers (like KFBK's Franklin array), top-loaded towers (like WHO in Des Moines), folded unipoles, and shunt-fed arrangements each solve different engineering problems. Modern Method of Moments (MoM) computer modeling has replaced the laborious field measurement proofs that once required engineers to drive to every tower and record readings by hand, but the fundamental physics, phased vertical radiators over a ground plane, hasn't changed since Ballantine's 1924 papers.
Chapter 2, Section 3, Part 2
Most AM directional arrays are built from towers of a single height. This is not an accident of construction; it is the easy path, because when every tower is the same electrical height, the current the phasor pushes into each one equals the field it radiates, and the licensing math stays simple. Depart from uniform height and that convenience breaks: a short tower and a tall tower fed the same current radiate different fields, so the current ratios and the field ratios diverge. The FCC's directional formula is written in field ratios precisely so it survives that divergence.
So the interesting question is not why arrays use equal heights. It is why some deliberately don't.
The FCC's licensed antenna database holds 3,872 current multi-tower AM directional systems. Of those, 87% use towers of a single electrical height. Only 497 mix heights at all. Set the bar at a 20-electrical-degree difference and 361 systems qualify (roughly 244 individual stations) and the population thins smoothly from there, with no natural break, down to two dozen arrays whose towers differ by more than 120 electrical degrees. Unequal height is the deliberate exception, chosen by about one array in eight.
| Height difference | Systems | Share of all multi-tower arrays | Stations | |
|---|---|---|---|---|
| any (all multi-tower) | 3,872 | 100% | 2,267 | |
| ≥ 5° | 497 | 12.8% | 334 | |
| ≥ 10° | 456 | 11.8% | 309 | |
| ≥ 20° | 361 | 9.3% | 244 | |
| ≥ 30° | 264 | 6.8% | 186 | |
| ≥ 40° | 210 | 5.4% | 150 | |
| ≥ 60° | 167 | 4.3% | 119 | |
| ≥ 80° | 125 | 3.2% | 92 | |
| ≥ 100° | 68 | 1.8% | 50 | |
| ≥ 120° | 24 | 0.6% | 17 | |
Table 2.3-1
A daytime pattern protects a bearing, hold the groundwave down toward the station you would otherwise interfere with. A nighttime pattern has a harder job: it must protect a bearing and an elevation angle, because after dark the signal that causes trouble is the skywave, leaving the towers at roughly 25-55° above the horizon and bouncing off the ionosphere into a distant co-channel station. Phasing (current ratio and timing) steers the null in azimuth. It cannot, on its own, control elevation. Tower height can. A tower's electrical height sets the shape of its vertical pattern, so building towers of different heights gives the designer a handle on the vertical plane that a uniform array simply does not have.
Drawing 2.3-2
That handle comes in two strengths.
The gentle version is a trim: two unequal towers roll off at different rates as you look upward, so the amplitude balance the designer struck at the horizon drifts with elevation, and a careful designer exploits the drift to deepen the null exactly where the skywave leaves. The dramatic version needs a tower taller than a half-wave (180 electrical degrees): such a tower has a genuine null in its own vertical pattern, an angle at which it radiates nothing at all, in every compass direction at once. Pick the height and you place that null on the skywave. The classic case is the 5/8-wave tower (about 225°), which maximizes groundwave along the horizon and, by the same physics, seen twice, nulls the sky near 37°. It is why 5/8-wave is called an anti-fading height.
KVXR in Moorhead, Minnesota runs a three-tower night array on 1,280 kHz with towers of 95°, 206°, and 70°. The middle tower is the tall one, past half-wave, with a vertical-pattern null around 48°. Counterintuitively it carries the least current in the array: a near-⅝-wave tower is such an efficient horizon radiator that a small current makes a respectable field. The short 70° tower is the opposite: it must be driven hard to contribute at all. Two towers, same array, at opposite ends of the height-versus-current trade.
These are pulled from the licensed database and span the full range: from a modest 53° trim to a tower more than a half-wave taller than its shortest neighbor. Every figure below is straight off the FCC record; the last column simply flags whether any tower exceeds 180 electrical degrees, the threshold above which a tower carries a null in its own vertical pattern.
| Station | City | kHz | Twr | Heights (°) | Spread | >180° |
|---|---|---|---|---|---|---|
| WPMZ | East Providence, RI | 1110 | 2 | 73-222 | 149° | yes |
| WOC | Davenport, IA | 1420 | 3 | 187-325 | 138° | yes |
| KVXR | Moorhead, MN | 1280 | 3 | 70-206 | 136° | yes |
| WADO | New York, NY | 1280 | 4 | 70-199 | 129° | yes |
| WNIO | Youngstown, OH | 1390 | 6 | 106-220 | 114° | yes |
| KAHZ | Pomona, CA | 1600 | 2 | 94-146 | 53° | no |
Table 2.3-3
A few are worth a second look. WOC's tallest tower runs 325 electrical degrees: far past the 5/8-wave optimum, into the region where the vertical pattern breaks into multiple lobes, and the point where the simple thin-wire model gets least trustworthy. WNIO stacks six towers, five of them over 180°, to shape a demanding night pattern. And KAHZ, the shortest-spread case, shows the gentle end of the idea: no tower has a null of its own, yet the two unequal heights still drift apart with elevation enough to be useful. The dramatic and the subtle are the same design, dialed to different intensities.
The first deliberate use of tower height to steer the skywave was WLW's 1935 fix. Running 500 kW on 700 kHz, the Cincinnati station was obliterating CFRB in Toronto. Crosley's engineers analyzed some twenty options and settled on two 326-foot "suppressor" towers, placed on the bearing toward Toronto and sized to knock the skywave down at about 20° above the horizon, a null aimed at a specific bearing and a specific elevation angle. Ninety years later, the same physics is licensed into 244 stations across the country.
A note on the word "suppressor": it was WLW's own 1935 label and does not survive as standard terminology; modern engineering says reference tower, field ratio, and phase. And two contemporary accounts disagree on whether the suppressor towers were driven or passive; both are on record, and the disagreement is itself a nice piece of history.
| Claim | Source |
|---|---|
| Current ratio ≠ field ratio for unequal heights; FCC formula written in field ratios | W.C. Alexander, Dir. of Engineering, Crawford Broadcasting: AM Antenna Systems |
| Night patterns protect a bearing and an elevation angle; height governs takeoff angle | Consulting engineer, RadioDiscussions multi-tower thread |
| Census: 3,872 multi-tower systems; 497 mixed-height; 361 at ≥20°; 244 stations; threshold curve | FCC CDBS licensed-antenna database (am_ant_sys, am_towers), January 2024 snapshot |
| Six-station parameters (heights, tower counts, field ratios, spreads) | FCC license records via fccinfo.com: WPMZ (fac. 4377), WOC (60360), KVXR (35863), WADO (70684), WNIO (13669), KAHZ (61814) |
| WLW suppressor towers: 326 ft, 1,850 ft out, nulling ~20° elevation toward Toronto (1935) | J. Schneider, The Development of the Directional AM Broadcast Antenna, Radio World |
| WLW suppressor towers reported as installed to reduce radiation toward Toronto (1935) | J. Brittain, Scanning the Past, Proceedings of the IEEE (1994) |
| WLW suppressor system eliminated CFRB interference without reducing signal strength | Edwin B. Dooley WLW Collection finding aid, University of Maryland Libraries |
| WLW towers fed via trolley wire, 85 kW at 96° out of phase (driven account; conflicts with passive account) | Wikipedia, WLW, citing period sources |
| Element-plane null formula (sin θ = 360/G − 1) and horizon field factor (1 − cos G) | Derived from standard thin-wire monopole theory; verified numerically |
| 36 electrical degrees = FCC threshold for directional-array protection | 47 CFR 1.30002 |
Table 2.3-4
AM Radio Facts and Trivia, the main page, with hundreds more records, oddities and firsts. Test yourself with the AM Trivia Challenge.
All AM radio stories by subject · AM radio interactive tools and simulators · AM Trivia Challenge quiz
© 2026 Chris, N6JET. Free to read and share for non-commercial use with credit (Creative Commons BY-NC 4.0). Not for sale.